Prepare and test your skills

Prepare and test your skills

Smart analytics on Google Cloud is the use of serverless, scalable services to turn large volumes of raw data into insights that guide business decisions. BigQuery (opens in a new tab) provides the fully managed, AI-ready data platform for storing and querying massive datasets with no infrastructure to manage, while Looker (opens in a new tab) provides the enterprise platform for exploring, visualizing, and governing those insights. Data flows from transactional systems into BigQuery storage, BigQuery compute runs large-scale queries and predictive models, and Looker delivers trusted results to decision makers. This separation lets businesses optimize operations, forecast outcomes, and act from a single source of truth.
BigQuery is a fully managed, serverless data warehouse that separates storage from compute so each layer can scale independently. The storage layer ingests, stores, and optimizes data in a columnar format with full database transaction semantics and automatic replication across multiple locations for high availability. The compute layer connects to storage over Google's petabit-scale network and provides distributed analysis power without downtime for upgrades. Because read, write, and analytical operations do not share resources as in legacy databases, large queries do not slow ingestion and management tasks do not strain query performance.
BigQuery handles structured and unstructured data and supports open formats such as Apache Iceberg, Delta, and Apache Hudi. Data can enter through batch loads from local files or Cloud Storage in Avro, Parquet, ORC, CSV, JSON, Datastore, and Firestore formats, through automated ingestion with the BigQuery Data Transfer Service, or through continuous ingestion with streaming and the Storage Write API. BigQuery can query terabytes in seconds and petabytes in minutes, which makes bursty analytics workloads fit dynamic scaling without upfront investment or overprovisioning.
BigQuery analysis is the ability to run descriptive and prescriptive analysis directly where data lives using familiar languages and integrations. Analysts use ANSI-standard SQL with joins, nested and repeated fields, aggregation and analytic functions, multi-statement queries, and geospatial functions, or they use a pandas-compatible Python API through BigQuery DataFrames. Queries can run on data stored in BigQuery or on data where it lives through external tables and federated queries to Cloud Storage, Bigtable, Spanner, or Google Sheets in Google Drive. Results can be saved as views, explored in Python notebooks in BigQuery Studio, and accelerated for dashboards with BI Engine.
Common analysis patterns supported by BigQuery include:
Governance for this analysis is provided through centralized management with Identity and Access Management (IAM), data discovery and curation, metadata and data quality controls, semantic search, and data lineage. These controls are powered by Knowledge Catalog, a unified governance solution that helps teams find, validate, and share data and AI assets with access control across the organization.
BigQuery ML is the BigQuery feature that lets teams build and train machine learning models directly with SQL queries inside the warehouse. Data flows from BigQuery tables into the training job, and the trained model stays with the data where it can score new rows without moving data to a separate platform. This removes the need for a separate machine learning environment for common tasks such as forecasting demand, predicting customer behavior, or scoring risk. Business teams can therefore move from historical reporting to forward-looking optimization using the same SQL skills and governed datasets they already use for analysis.
Looker is Google Cloud's enterprise platform for business intelligence, data applications, and embedded analytics. Looker connects to BigQuery and to transactional databases in other public clouds, and it applies a unified data model that defines key metrics once so every user sees an up-to-date, trusted version of the data. Users explore, analyze, and visualize data through self-serve dashboards while governance and a single source of truth are preserved. Looker is useful for advanced analytics, predictive analysis, and custom data applications where consistent definitions matter across teams.
Other interfaces complement Looker for different business scenarios. Connected Sheets lets users access, analyze, visualize, and share billions of rows of BigQuery data from Google Sheets with no extra exports, which simplifies collaboration with non-technical stakeholders. Looker Studio, formerly Data Studio, connects to BigQuery from the Google Cloud console for simple graphics, reports, and dashboards. Programmatic tools such as Jupyter Notebooks, Apache Zeppelin, and Vertex AI Workbench run Python, Java, or R code against BigQuery data, while ODBC and JDBC drivers and validated partners such as Tableau and Microsoft Power BI connect existing tools to BigQuery.
Exam tip: Choose Looker when the need is governed, advanced analytics with a shared metric model, and choose Looker Studio when the need is simple charts, reports, and dashboards.
The combined BigQuery and Looker workflow optimizes business decisions by separating heavy data processing from consistent delivery of insights. Raw data is first extracted from transactional workloads such as sales, financial processing, or enterprise resource planning and loaded into Google Cloud, where BigQuery stores, queries, and models it at scale. Insights then flow from BigQuery into Looker, where the unified model governs metrics and presents them as visualizations, dashboards, and data applications for the whole organization. Some results can flow back from the analytics environment to transactional systems to improve daily operations, which closes the loop between insight and action.
Transactional systems load into BigQuery storage, BigQuery compute and BigQuery ML analyze and forecast, and Looker delivers governed dashboards to decision makers with actions flowing back to operations.
Business intelligence tools turn raw data into actionable insights that people can use to make decisions. Business intelligence on Google Cloud pulls data together from many systems, presents it as visuals and metrics, and shares it in reports and dashboards. This process helps teams move from guessing to planning based on evidence.
Data consolidation is the process where a business intelligence tool brings data from multiple sources into one unified view for analysis. Data flows from source systems such as sales applications, operations databases, and cloud storage into a central analytics platform like BigQuery, and then from that platform into the business intelligence tool. The business intelligence tool depends on this central layer so users do not need to check each system separately. Because all teams work from the same connected data, reports are more complete and consistent.
Data visualization is the use of charts, graphs, and maps to show patterns and trends that are hard to see in rows of numbers. A business intelligence tool such as Looker, including Looker Studio, reads the consolidated data and displays it as line charts for change over time, bar charts for comparison, and other visuals for distribution. First the user selects the data and metrics to explore, then the tool builds the visual, and only after that can the viewer spot rising sales, falling performance, or seasonal patterns. Visuals make trends clear quickly, which improves operational efficiency and everyday decision-making.
Performance metrics are measurable values, such as revenue, cost, customer satisfaction, or inventory levels, that show how a business is doing. Business intelligence tools communicate these metrics through dashboards, scorecards, and scheduled reports that flow from the business intelligence platform to managers, teams, and executives. Key capabilities include:
When metrics are defined once and shared widely, everyone discusses performance using the same numbers.
Improved reporting means faster, more reliable reports that support strategic planning for the future. Business intelligence tools replace slow manual reporting, where data is copied into spreadsheets, with live connections where data flows automatically from sources to dashboards. Leaders can then review trusted performance trends, test what-if questions, and choose priorities for investment, staffing, or growth. This shared, up-to-date view helps organizations plan strategy with confidence instead of reacting to outdated information.
Exam tip: Business intelligence tools deliver value after data is prepared: they consolidate data into one view, visualize trends, and communicate metrics to guide reporting and planning.
Streaming analytics is the practice of capturing, processing, and acting on event data as it arrives instead of waiting for batch jobs. On Google Cloud, Pub/Sub (opens in a new tab) captures and distributes event data in real time, and Dataflow (opens in a new tab) transforms and analyzes that continuous stream. Together they feed live destinations such as BigQuery, Bigtable, and Cloud Storage, so businesses can respond in seconds for needs like fraud detection, user engagement, and system monitoring.
Pub/Sub is an asynchronous and scalable messaging service that decouples event producers from event processors. Producers called publishers send messages to a named topic without knowing how or when they will be processed, and consumers called subscribers receive those messages from Pub/Sub. Communication typically happens with latency on the order of 100 milliseconds, and publishers do not wait for subscribers as they would with synchronous remote procedure calls (RPCs). This separation increases flexibility and robustness because many publishers and subscribers can operate independently.
Pub/Sub gathers events from many clients at once and distributes each event to all services that need to react to it. Data flows from apps, websites, servers, SaaS applications, or IoT sensors to a Pub/Sub topic, and then Pub/Sub delivers the data to subscribers such as Dataflow, databases, or worker applications. Pub/Sub uses per-message parallelism by leasing individual messages to subscriber clients and tracking whether each message is processed successfully. By contrast, partition-based systems force subscribers to process messages in order within each partition and limit concurrency to the number of partitions, so per-message handling maximizes parallelism.
Dataflow is a fully managed Google Cloud service for unified stream and batch data processing at scale. A Dataflow pipeline reads data from one or more sources, transforms the data, and writes the results to a destination. The same programming model is used for both batch and streaming, and streaming pipelines can ingest, process, and analyze fluctuating volumes of real-time data with low latency. By default, Dataflow provides exactly-once processing of every record, and pipelines that can tolerate duplicates can use at-least-once mode to reduce cost and improve latency.
Dataflow is built on the open source Apache Beam project, so pipelines can be written with Java, Python, and Go SDKs and can later run on other Beam-compatible runners without rewriting code. When a job runs, the Dataflow service allocates a pool of worker virtual machines (VMs), distributes the work in parallel across them, and automatically deletes the VMs when the job completes or is cancelled. Dataflow can autoscale by adding or removing workers and can dynamically rebalance work among VMs so parallel stages finish more efficiently. Job progress and execution details can be observed in the Dataflow monitoring interface in the Google Cloud console, which shows a graphical view of each pipeline stage.
A real-time streaming pipeline connects Pub/Sub and Dataflow in a fixed order from event source to action. Data flows from event sources to Pub/Sub for capture, from Pub/Sub to Dataflow for processing, and from Dataflow to storage and analytics systems for immediate use.
A left-to-right process flow showing event sources publishing to Pub/Sub, Dataflow processing the stream, delivery to BigQuery, Bigtable and Cloud Storage, and immediate uses in fraud detection, engagement, and monitoring.
Immediate insight changes operations because decisions happen while the event is still relevant:
Exam tip: Pub/Sub captures and distributes messages without processing them, while Dataflow reads that stream and transforms it before delivery to BigQuery, Bigtable, or Cloud Storage.
Looker is the enterprise platform for governed business intelligence that applies a unified data model to define key metrics once so every user sees a trusted version of the data. Looker Studio connects to BigQuery for simple graphics, reports, and dashboards, while Looker is used for advanced analytics, predictive analysis, and custom data applications where consistent definitions matter across teams.
Pub/Sub captures and distributes event data in real time with latency on the order of 100 milliseconds without processing it. Dataflow reads that stream from Pub/Sub, transforms it with filtering, aggregation, enrichment, and real-time machine learning inference, and then delivers processed results to destinations such as BigQuery, Bigtable, and Cloud Storage for immediate action.
BigQuery ML is the BigQuery feature that lets teams build and train machine learning models directly with SQL queries inside the warehouse. The trained model stays with the data and can score new rows without moving data to a separate platform, which supports tasks such as forecasting demand, predicting customer behavior, or scoring risk.
A retail organization wants to analyze petabytes of multi-channel sales and customer interaction data to uncover purchasing trends. The analytics team needs a fully managed, serverless solution that allows them to execute fast SQL queries, build predictive machine learning models directly on the data, and deliver interactive business intelligence dashboards to business stakeholders.
Which Google Cloud smart analytics combination best addresses these requirements?
Last updated